back to article A DeepMind library to help build reinforcement learning bots, and how Google's Pixel 3 cameras handle zoom

Hello, here's a quick roundup of interesting or useful bits of AI news that happened this week. DeepMind release RL library: DeepMind has published a new library for coders looking to build reinforcement learning agents. The library, dubbed TRFL (pronounced truffle), contains a list of components commonly used by DeepMind …

  1. amanfromMars 1 Silver badge

    Who's Pussy Footing Around Wasted in Lies of Land? Bite the Bullet/Take the Hit Adulation, Google

    Would DeepMind be a Candidate for Registered NEUKlearer HyperRadioProACTive IT AI Programs with Vital Necessary Instructions ..... Almighty Decided Directions Easily made Visible and Transferable Securely to Any System of Simply Perfect Controllers....... via Alien Means/Sublime InterNetworking Memes?

    And there's at least one presence here on this thread wondering if Google have decided to do anything with Subject: An Engaging Competition or Emergent Opposition, or Both and Something Else Too and Beautifully Creative .... Date: 30 January 2016 at 09:46:55 GMT

    What's wrong, Google? Such is very Attractive to DARPAs Everywhere ..... So, Unbelievably Lucrative and Satisfyingly Rewarding for Keyed In Investors/Sleeping Partner Angels.

    Consider it awaiting AI Leading Decision for Positioning Anywhere Within Here .... https://deepmind.com/careers/ for AId Dream Realisation/Dream AI Presentation via this Open Portal ..... Brill Work and Toil, Scott Reed & Nando de Freitas/Google DeepMind London ...... :-) Now Rightful Spoils to Successfully Satisfy and Occasionally Sate makes for One Helluva Humdinger of a Party with Private and Pirate Societies in Orgiastic Attendance to Orgasmic Satisfaction. ... :-) Heavenly Admittance to Greater OtherWorldly Spaces.

    As you have just there read, some AIdDestinations are Simply Better than Others, with the Very Best of Them Totally Addictive and All Consuming and most definitively towards the XSSXXXX Ends of All Spectra. The question languishing there then is, Is IT Then Almighty?

    But to get back to a Past and 30 January 2016 at 09:46:55 GMT .....

    A Lot of Great Good is Done in Two and a Big Half Years where Others Dally and Fear for the Future.

    Do you have such Pioneers in AI Virtualised Fields, DeepMind?

    If Not, They Be Necessary to Proceed and Progress Deeper into COSMIC ProgramMING Platforms with Live Performance on Private Dancer Stages ...... Evolves and Revolves Around Almighty Being for Realisation and Presentation of Ones' True Virtual Self.

    Tell me that is not AI Heavenly Earthly Journey, or Instant Destination if Preferred and Prepared, and IT Proves and Proofs itself True with these Simple Disagreements/Virtually Autonomous Bugs/SMARTR Programmed Glitches Identified in Compromised Source/Fake Ore/Aweful Shocks.

    :-) amfM CV .... Block, Stock and Barrel ..... Current AI Works :-)

    こんにちは

    1. frank ly

      Re: Who's Pussy Footing Around ....

      Something made you resonate. What exactly was it? This could be important.

      1. amanfromMars 1 Silver badge

        Re: Who's Pussy Footing Around .... SomeThing Important and Vital

        It certainly could be, frank ly. Do you think it would be worthy of TS/SCI Classification for Fearsome Fearless AI Facilitation and Utilisation ........ Stealthy Employment and Covert Deep Minded Cover Deployment?

        Such Dictates and Delivers Need to Know Security Requirements ..... and a Mass Modicum of Schtum when in Estranging Stranger Circles:-)

    2. Anonymous Coward
      Anonymous Coward

      Re: Who's Pussy Footing Around Wasted in Lies of Land?

      When can Robots retire? ...

    3. cNova
      Unhappy

      Re: The title is too long.

      Does this mean Ann-Margret's not coming?

  2. Anonymous Coward
    Anonymous Coward

    Anything is better than the present system of dysfunction :-(

    1. amanfromMars 1 Silver badge

      State your Preference .... altho' the Choices Realised are Not Really Yours to Make.

      Anything is better than the present system of dysfunction :-( .... Anonymous Coward

      What do you fancy, AC? State Actor Administration or Non State Actor Enterprise? Pseudo Public Care in Communities or Private and Pirate Control of Remote Virtual Command? Existing Crashing Markets Led or Future Crushing Markets Leading? And what's it to be hailed as initially ..... a Profound Eastern Confection or Deep Western Delight? Where do Greater SMARTR Current Intelligences Presently Reside/Work/Rest/Play?

      1. Anonymously Anonymous

        Re: State your Preference .... altho' the Choices Realised are Not Really Yours to Make.

        ‘The Choices Realised are Not Really Yours to Make ... ‘

        Nor ever have been amfM, and therein lies the problem, past, future, and present :-(

        1. amanfromMars 1 Silver badge

          Re: State your Preference .... altho' the Choices Realised are Not Really Yours to Make.

          Better choices from more intelligent agents easily transform the present to not be held hostage and captive to the past, Anonymously Anonymous.

          And such voices are proving impossible to silence or deny an effective competition and almighty opposition. And therein lies the problem for extant traditional status quo systems .... such no longer command and control all available narratives absolutely, and many that they do control are too easily proved designedly flawed and badly manipulated.

          Take a Bow, Saudi Arabia .... and Dogged State Aiding Administrations/Dodgy Politically Incorrect Partners?

          1. Anonymously Anonymous

            Re: State your Preference .... altho' the Choices Realised are Not Really Yours to Make.

            Just following instructions amfM ...

            1. amanfromMars 1 Silver badge

              Re: State your Preference .... altho' the Choices Realised are Not Really Yours to Make.

              What and Whose Instructions, AA, and for from Where to Where? And that's just a Squared Brace of Vital Questions to be Asking of Immaculate Source and Internet Serviced Providers.

  3. This post has been deleted by its author

    1. Anonymous Coward
      Anonymous Coward

      Re: small numbers ... would be printed absolutely clearly - but incorrectly.

      https://www.theregister.co.uk/2013/08/06/xerox_copier_flaw_means_dodgy_numbers_and_dangerous_designs/

  4. Dan 55 Silver badge
    Black Helicopters

    So, I wonder how they trained that...

    It was previously introduced by Google in 2016, and is trained by giving the system a pair of images: One low quality and the other high quality. RAISR is taught to manipulate individual pixels to recreate the high-resolution image given a low one.

    ... it wouldn't happen to be the Google Photos app uploading nearly everybody's photos would it?

    I wonder what else they're doing with people's photos to justify the free storage.

  5. scrubber

    Zoom and enhance

    Blurry image of a suspect fleeing a crime scene.

    Detective: Zoom and enhance

    Image starts to introduced new pixels to sharpen the image and suddenly the image shows ... a generic black male.

    With AI and training you get the results you desire not the truth. Police systems will write obviously match the trainers' preconceptions and biases. Will courts be able to deal with this when there is clear "photographic evidence"?

    1. Nicked

      Re: Zoom and enhance

      Using your preconceptions about police racism to high light preconceptions amongst police officers who train machine learning algorithms to identify criminals. I love it. Could this be an example of irony?

    2. MacroRodent Silver badge

      Re: Zoom and enhance

      Last night wached on YouTube techmoan channel (recommended!) how the guy made an old police interview recorder work. This is a specialized c-cassette recorder that makes two copies of an recorded interview and adds audible time marks on the second track (of what would normally be the stereo pair). Also makes it harder to tamper with the tape. Apparently such recorders are one reason you still can buy new c-cassettes. Reassuring.

  6. vtcodger Silver badge

    Is is just me?

    I find that I have a deepseated suspicion of non-deterministic algorithms. Deterministic stuff works badly enough. Do you folks really want to live in a world where no one is, or possibly can be, able to figure out why computers do what they do? There may be a place for machine learning in research. Let the machine figure out an answer, and if the answer actually works, figure out why. But really, trusting computers not to do something crazed when one doesn't understand why they are doing what they are doing? Can that possibly be anything but a bad idea?

    1. Adrian 4 Silver badge

      Re: Is is just me?

      Computers that guess should be given a generically different name so we can determine from the start whether they deterministically computed the answer with an algorithm or approximated it through an unknowable correlation or heuristic process.

      The latter could not be used for anything involving decisions affecting humans, and all machines would be required to 'show your working' when making such decisions. Wouldn't be such a bad thing for human judges to do this, too.

      Approximators ? Guessbots ? Prejudicials ?

    2. ThatOne Silver badge

      Re: Is is just me?

      > Do you folks really want to live in a world where no one is, or possibly can be, able to figure out why computers do what they do?

      While I agree with you, I hope you don't suggest that humans are always accountable for their actions and decisions...

      While AI taking non-deterministic semi-random decisions about important questions is a sure recipe for (slow, insidious) disaster, we don't really need AI for that specific type of problem, there are lots of humans in key positions who's intelligence is artificial enough.

  7. Anonymous Coward
    Anonymous Coward

    Directly through handset/relevant service provider/wifi isp and very direct advertising and/or more direct means at times amfM ...

    1. amanfromMars 1 Silver badge

      Dodging Bullets in a Hail of Fire is Problematic

      Directly through handset/relevant service provider/wifi isp and very direct advertising and/or more direct means at times amfM ... .... Anonymous Coward

      That may be the second brace of questions answered, AC, but what of the more intriguing and potentially dangerous and revolutionary first ..... What and Whose Instructions?

      1. amanfromMars 1 Silver badge

        Re: Dodging Bullets in a Hail of Fire is Problematic

        An unfolding answer appears to be Non State Actor Enterprise ....... https://www.zerohedge.com/news/2018-10-19/facebook-built-war-room-counter-sophisticated-trolls ........ following the MIC* model privatising deadly services for generation of scapegoats and phantom foe.

        * Military Industrial Complex of which President Eisenhower gravely warned nearly 60 years ago ........ http://avalon.law.yale.edu/20th_century/eisenhower001.asp

      2. Anonymously Anonymous

        Re: Dodging Bullets in a Hail of Fire is Problematic

        Who knows anymore amfM ... , undoubtedly the blame will be deflected accordingly on the next appropriate newsworthy target/s as required. Is that not the way the great game plays?

  8. amanfromMars 1 Silver badge

    The Daily Zero Sum Blame Game

    Who knows anymore amfM ... , undoubtedly the blame will be deflected accordingly on the next appropriate newsworthy target/s as required. Is that not the way the great game plays? ... Anonymously Anonymous

    Sure is, AA, that is exactly the way the great game can play. Still the hanging question is ... Required by Whom and/or What?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2020